Search

DNA methylation has a role in mediating epigenetic silencing of CpG island genes in cancer and other diseases. Identification of all gene promoters methylated in cancer cells "the cancer methylome" would greatly advance our understanding of gene regulatory networks in tumorigenesis. We previously described a new method of identifying methylated tumor suppressor genes based on pharmacologic unmasking of the promoter region and detection of re-expression on microarray analysis. In this study, we modified and greatly improved the selection of candidates based on new promoter structure algorithm and microarray data generated from 20 cancer cell lines of 5 major cancer types. We identified a set of 200 candidate genes that cluster throughout the genome of which 25 were previously reported as harboring cancer-specific promoter methylation. The remaining 175 genes were tested for promoter methylation by bisulfite sequencing or methylation-specific PCR (MSP). Eighty-two of 175 (47%) genes were found to be methylated in cell lines, and 53 of these 82 genes (65%) were methylated in primary tumor tissues. From these 53 genes, cancer-specific methylation was identified in 28 genes (28 of 53; 53%). Furthermore, we tested 8 of the 28 newly identified cancer-specific methylated genes with quantitative MSP in a panel of 300 primary tumors representing 13 types of cancer. We found cancer-specific methylation of at least one gene with high frequency in all cancer types. Identification of a large number of genes with cancer-specific methylation provides new targets for diagnostic and therapeutic intervention, and opens fertile avenues for basic research in tumor biology.

Deleted in colorectal cancer (DCC) is a candidate tumor-suppressor gene located at chromosome 18q21. However, DCC gene was found to have few somatic mutations and the heterozygous mice (DCC(+/-)) showed a similar frequency of tumor formation compared with the wild-type mice (DCC(+/+)). Recently, DCC came back to the spotlight as a better understating of its function and relationship with its ligand (netrin-1) had shown that DCC may act as a conditional tumor-suppressor gene. We evaluated hypermethylation as a mechanism for DCC inactivation in head and neck squamous cell carcinoma (HNSCC). DCC promoter region hypermethylation was found in 75% of primary HNSCC. There was a significant correlation between DCC promoter region hypermethylation and DCC expression (assessed by immunohistochemistry; P = 0.021). DCC nonexpressing HNSCC cell lines JHU-O12 and JHU-O19 with baseline hypermethylation of the DCC promoter were treated with 5-aza-2'-deoxycytidine (a demethylating agent) and reexpression of DCC was noted. Transfection of DCC into DCC-negative HNSCC cell lines resulted in complete abrogation of growth in all cell lines, whereas additional cotransfection of netrin-1 resulted in rescue of DCC-mediated growth inhibition. These results suggest that DCC is a putative conditional tumor-suppressor gene that is epigenetically inactivated by promoter hypermethylation in a majority of HNSCC.

We showed that the PEA3 transcriptional factor interacted with LKB1, a serine/threonine kinase, which is somatically mutated in lung cancer. This interaction occurred through the ETS domain of PEA3 and the kinase domain of LKB1. Mutation of LKB1 in lung cancer cells stabilized PEA3. Reintroduction of wild-type (WT) LKB1 into cells induced down-regulation of PEA3 and subsequently resulted in reduced cyclooxygenase-2 RNA and protein expression, whereas germ-line and somatic LKB1 mutants were defective in this activity. LKB1 phosphorylated PEA3 and promoted its degradation through a proteasome-mediated mechanism. Cells expressing mutant LKB1 possessed greater invasive potential compared with cells expressing WT LKB1. Increased invasion of cells with mutant LKB1 was partly due to PEA3 expression, as RNA interference inhibition of PEA3 resulted in dramatic decrease of Matrigel invasion. However, forced expression of PEA3 resulted in down-regulation of epithelial markers and induction of mesenchymal markers. These results suggest that PEA3 stabilization due to LKB1 inactivation could lead to epithelial/mesenchymal transition and greater lung cancer invasion potential.

Epidemiological studies have demonstrated a causal association between tobacco use and carcinoma of the lung, and some genetic targets of the carcinogens in cigarette smoke have been defined recently. We further examined the effect of cigarette smoking on the frequency of allelic losses on chromosome 9p21 and the incidence of p16 inactivation. Chromosomal loss at 9p21-24 was determined by microsatellite analysis using 14 markers in 47 patients with non-small cell lung cancer. In addition, p16 gene inactivation was determined by DNA sequence analysis, methylation-specific PCR, and immunohistochemistry. Tumors from a group of nonsmokers (n = 14) were compared with tumors from a group of smokers (n = 33) matched for cell type, tumor stage, and gender. Allelic loss encompassing the p16 locus was present significantly (P = 0.01) more often in smokers (23 of 33 smokers, 70%) than in nonsmokers (4 of 14 nonsmokers, 28%). No significant differences in the frequency of p16 inactivation were observed between smokers and nonsmokers (45% versus 36%). However, homozygous deletion of the p16 gene locus and point mutation of p16 gene were only observed in tumors from smokers, whereas the p16 gene was inactivated in tumors from nonsmokers only through promoter hypermethylation. Thus, inactivation of the p16 gene is a common event in all non-small cell lung cancer, but the mechanism of gene alteration differs between smokers and nonsmokers. The significant link between tobacco and loss of the p16 locus identifies additional genetic targets of smoking in the pathogenesis of lung cancer.

The etiology of lung tumors arising in nonsmokers remains unclear. Although mutations in the K-ras and p53 genes have been reported to be significantly higher in smoking-related lung carcinomas, in the present study we performed a more comprehensive analysis in search of additional genetic changes between lung adenocarcinoma from tobacco- and non-tobacco-exposed patients. We selected a matched cohort of 18 lifetime nonsmoking and 27 smoking patients diagnosed with primary adenocarcinoma of the lung and searched for chromosomal alterations in each tumor by testing normal and tumor tissue with 54 highly polymorphic microsatellite markers located on 28 different chromosomal arms. Allelic losses or gains at chromosomal arms 3p (37 versus 6%), 6q (46 versus 12%), 9p (65 versus 22%), 16p (28 versus 0%), 17p (45 versus 11%), and 19p (58 versus 16%) were present significantly more often in adenocarcinomas from smokers than from nonsmokers. Chromosomal arms showing allelic imbalance in lung tumors from nonsmokers were rare but occurred more often at 19q (22%), 12p (22%), and 9p (22%). The FAL (fractional allelic loss or gain) is defined as the percentage of chromosomal arm losses/gains among the total informative chromosomal arms. Tumors from smokers harbored higher levels of FAL (13 (48%) of 27 showed FAL > or = 0.3) compared with the lung tumors from the nonsmoker patients (2 (11%) of 18 showed FAL > or = 0.3; P = 0.02; odds ratio, 0.13; 95% confidence interval, 0.01-0.79). Our data demonstrate that widespread chromosomal abnormalities are frequent in lung adenocarcinoma from smokers, whereas these abnormalities are infrequent in such tumors arising in nonsmokers. These observations support the notion that lung cancers in nonsmokers arise through genetic alterations distinct from the common events observed in tumors from smokers.

Aberrant promoter hypermethylation is common in head and neck cancer and may be useful as a marker for cancer cells. We examined whether cells with tumor-specific aberrant DNA-methylation might be found in the saliva of affected patients. We tested 30 patients with primary head and neck tumors using methylation-specific PCR searching for promoter hypermethylation of the tumor suppressor gene p16 (CDKN2A), the DNA repair gene O6-methylguanine-DNA-methyltransferase (MGMT) and the putative metastasis suppressor gene death-associated protein kinase (DAP-K). Aberrant methylation of at least one of these genes was detected in 17 (56%) of 30 head and neck primary tumors; 14 (47%) of 30 at p16, 10 (33%) of 30 at Dap-K and 7 (23%) of 30 at MGMT. In 11 (65%) of 17 methylated primary tumors abnormal methylated DNA was detected in the matched saliva samples. Abnormal promoter methylation in saliva DNA was found in all tumor stages and more frequently in tumors located in the oral cavity. Moreover, none of the saliva from patients with methylation-negative tumors displayed methylation of any marker. Of 30 saliva samples from healthy control subjects (15 smokers and 15 nonsmokers), only one sample from a smoking patient was positive for DNA methylation at two target genes. Detection of aberrant promoter hypermethylation patterns of cancer-related genes in saliva of head and cancer patients is feasible and may be potentially useful for detecting and monitoring disease recurrence. Long-term longitudinal studies are needed to evaluate this approach for early detection of head and neck cancer in at-risk populations.

Genetic alterations of chromosome 7 are common in human cancer. Furthermore, previous studies have supported the presence of a gene important in a broad range of cancers at 7q22-31.1. There is evidence that supports an oncogenic function for this putative gene, as well as evidence that supports a tumor suppressive role. In this study, we used a cross-species candidate gene approach in combination with physical mapping to identify MPP11 as a candidate for the putative cancer-related activity at 7q22-31.1. We then analyzed primary head and neck squamous cell tumors (HNSCCs) for loss of heterozygosity/allelic imbalance (LOH/AI) at the MPP11 genomic locus. Thirty-eight percent of tumors examined displayed LOH/AI involving the MPP11 genomic locus. Mutation analysis of MPP11 in the latter samples did not identify any inactivating mutations. However, immunohistochemical staining of primary tumor sections and Western blot analysis of HNSCC cell lines revealed a tumor-specific high level of expression of MPP11p. Fluorescence in situ hybridization analysis done on the cell lines identified increased chromosome 7 copy number with a concomitant increase in MPP11 copy number. These results suggest an oncogenic role for MPP11 in HNSCC.

Recently, mutations of the transforming growth factor-beta receptor type I gene have been reported to occur at high frequency in breast cancer metastases, with all mutations being an identical C to A transversion at nucleotide 1160 of the gene (T. Chen et al, Cancer Res., 58: 4805-4810, 1998). This mutation would result in a serine to tyrosine substitution at codon 387 (S387Y) and would reportedly disrupt receptor function. Because this mutation reportedly occurred at high frequency in breast cancer metastases (42%) and much less frequently in primary breast cancer tumors (6%), this would seem to represent a pivotal genetic alteration in breast cancer progression. To further investigate the possible role of this specific genetic alteration in the progression of breast cancer and other forms of adenocarcinoma, we analyzed 20 breast cancer metastases, 15 lung adenocarcinoma metastases, and 13 colorectal cancer metastases for possible mutations at this site. Using both single-strand conformation polymorphism screening and sequencing, we found no mutations of this gene in any of our samples. Our results suggest the S387Y mutation of the transforming growth factor-beta receptor type I gene is not common in these types of human cancers.

We used the serial analysis of gene expression (SAGE) method to systematically analyze transcripts present in non-small cell lung cancer. Over 226,000 SAGE tags were sequence analyzed from two independent primary lung cancers and two normal human bronchial/tracheal epithelial cell cultures. A total of 226,000 SAGE tags were sequence identified, representing 43,254 unique transcripts. Comparison of the tags present in the tumor with those identified in the normal tissue revealed 175 transcript tags that were overrepresented in the normal tissue and 142 tags that were overexpressed in the tumor by 10-fold or more. Northern hybridization was performed on 15 of the most abundantly expressed tags identified in the tumors. These tags were derived from either a known gene or a matched expressed sequence tag clone. The transcripts for 3 of the 15 genes, PGP 9.5, B-myb, and human mutT, were abundantly expressed in primary lung cancers (10 of 18, 15 of 18, and 6 of 12 tumors, respectively). In contrast, the presence of PGP9.5 and B-myb was much less frequent in primary tumors derived from other tissue origins. These results suggest that at least a portion of the transcripts identified by SAGE are frequently associated with lung cancer, and that their overexpression may contribute to lung tumorigenesis. The identification and further characterization of genes generated by SAGE should provide potential new targets for the diagnosis, prognosis, and therapy of lung cancer.

Expression of the invasion/metastasis suppressor, E-cadherin, is diminished or lost in thyroid carcinomas. Yet, mutational inactivation of E-cadherin is rare. Herein, we show that this loss is associated with hypermethylation of the E-cadherin 5' CpG island in a panel of human thyroid cancer cell lines. This aberrant methylation is evident in 83% of papillary thyroid carcinoma, 11% of follicular thyroid carcinoma, 40% of Hurthle's cell carcinoma, and 21% of poorly differentiated thyroid carcinomas. Contrary to previous reports, the majority of these poorly differentiated thyroid carcinomas express E-cadherin, but often within the cytoplasm rather than at the cell surface. Together, our data indicate that the invasion/metastasis suppressor function of E-cadherin is frequently compromised in human papillary, Hurthle's cell, and poorly differentiated thyroid carcinoma by epigenetic and biochemical events.

The inability to distinguish microinvasive follicular thyroid cancer from benign follicular tumors preoperatively presents an important surgical dilemma. We examined 44 follicular tumors and found telomerase activity in all 11 follicular carcinomas and in 8 of 33 benign follicular tumors. It was undetectable in 22 normal thyroid tissues adjacent to the tumors. Telomerase activity may thus provide a diagnostic marker distinguishing benign from malignant follicular thyroid tumors. The ability to identify invasive follicular thyroid tumors could avert over 14,000 thyroidectomies annually in the United States, thereby significantly decreasing morbidity and health care costs.

Telomerase is a ribonucleoprotein that maintains telomere length and whose activity is associated with escape from cellular senescence. Telomerase activity has been found in germline, immortalized, and malignant tumor cells. Using a modified PCR-based assay for telomerase activity, 26 of 35 (80%) primary, fresh, head and neck squamous cell cancer specimens and 3 of 6 head and neck squamous dysplastic lesions possessed telomerase activity. In addition, 14 of 44 (32%) oral rinses from a separate group of head and neck squamous cell cancer patients contained detectable telomerase activity, whereas 1 of 22 (5%) oral rinses from normal control patients exhibited telomerase activity. Telomerase activity in oral rinses was compared with corresponding activity in paired primary tumor samples for 19 cases: 7 of 19 demonstrated activity in both tumor and oral rinse, 2 of 19 lacked activity in both tumor and oral rinse, 10 of 19 tumors demonstrated activity that could not be detected in corresponding oral rinses, and there were no examples of positive oral rinses with corresponding negative tumors. Although currently limited in its sensitivity, analysis of telomerase activity in oral rinses represents a novel method to detect the presence of cancer cells shed in the upper aerodigestive tract.

The tumor suppressor gene p16 (CDKN2/MTS-1/INK4A) can be inactivated by multiple genetic mechanisms. We analyzed 29 invasive primary head and neck squamous cell carcinomas (HNSCC) for p16 inactivation with immunohistochemistry utilizing a new monoclonal antibody (mAb), DCS-50. p16 staining of the primary lesions was correlated with genetic analysis including: (a) detailed microsatellite analysis of markers at the p16 locus to detect homozygous deletion; (b) sequence analysis of p16; and (c) Southern blot analysis to determine the methylation status of the 5' CpG island of p16. Twenty-four of 29 (83%) head and neck squamous cell carcinoma tumors displayed an absence of p16 nuclear staining using immunohistochemistry. Of these 24 tumors, we found that 16 (67%) harbored homozygous deletions, 5 (21%) were methylated, 1 displayed a rearrangement at the p16 locus, and 1 displayed a frameshift mutation in exon 1. These data suggest that: (a) inactivation of the p16 tumor suppressor gene is a frequent event in squamous cell carcinomas of the head and neck; (b) p16 is inactivated by several distinct and exclusive events including homozygous deletion, point mutation, and promoter methylation; and (c) immunohistochemical analysis for expression of the p16 gene product is an accurate and relatively simple method for evaluating p16 gene inactivation.

A genetic progression model of head and neck squamous cell carcinoma has not yet been elucidated, and the genetic basis for "field cancerization" of the aerodigestive tract has also remained obscure. Eighty-seven lesions of the head and neck, including preinvasive lesions and benign lesions associated with carcinogen exposure, were tested using microsatellite analysis for allelic loss at 10 major chromosomal loci which have been defined previously. The spectrum of chromosomal loss progressively increased at each histopathological step from benign hyperplasia to dysplasia to carcinoma in situ to invasive cancer. Adjacent areas of tissue with different histopathological appearance shared common genetic changes, but the more histopathologically advanced areas exhibited additional genetic alterations. Abnormal mucosal cells surrounding preinvasive and microinvasive lesions shared common genetic alterations with those lesions and thus appear to arise from a single progenitor clone. Based on these findings, the local clinical phenomenon of field cancerization seems to involve the expansion and migration of clonally related preneoplastic cells.

Patients with head and neck cancers have a high (2-3%/year) incidence of second primary lesions. Clinically, these new lesions are identified either simultaneously with the primary lesion (synchronous) or after a period of time (metachronous). This observation has been attributed to the concept of "field carcinogenesis," which is based on the hypothesis that prolonged exposure to carcinogens leads to the independent transformation of multiple epithelial cells at several sites. An alternative theory is based on the premise that any transforming event is rare; following initial transformation, the progeny of the transformed clone spread through the mucosa and give rise to geographically distinct but genetically related tumors. We analyzed the pattern of X-chromosome inactivation in multiple primary tumors from eight female patients with head and neck cancer. In addition, we used microsatellite analysis to examine the pattern of allelic loss on chromosomes 9p and 3p, identified as early events in the progression of head and neck malignancies. In four of four cases, multiple tumors demonstrated the same pattern of X-chromosome inactivation. In the remaining four cases, X-chromosome deletions prevented interpretation (n = 3), or the androgen receptor locus was noninformative (n = 1). In three of nine patients, multiple tumors displayed the same pattern of loss of heterozygosity, two with identical breakpoints on chromosome 9p. In one patient, there was an identical microsatellite alteration at a 3p locus, definitive evidence that these tumors arose from the same clone. Our findings suggest that in at least a proportion of patients with head and neck cancers, multiple primary tumors arise from a single clone.

Atypical alveolar hyperplasia (AAH) is a potential precursor lesion from which lung adenocarcinomas arise and may be a good target for studying the early events of lung tumorigenesis. A common genetic alteration in lung adenocarcinomas is mutational activation of K-ras. To determine the timing of K-ras activation, we evaluated formalin-fixed and paraffin-embedded tissue samples of 41 AAHs and their paired lung neoplasms from 28 patients for codon 12 point mutations of the K-ras oncogene. K-ras codon 12 mutations were detected using PCR followed by allele-specific oligonucleotide hybridization. Mutations were found in 16 (39%) of the 41 AAHs, 8 (42%) of the 18 adenocarcinomas, and none (0%) of the 5 lung neoplasms that were not adenocarcinomas. Of the 18 patients with both an AAH and a synchronous lung adenocarcinoma, 6 had K-ras mutation in the adenocarcinoma but not in the AAH, 6 had mutations in the AAH but not in the adenocarcinoma, 4 did not harbor mutations in either the AAH or the adenocarcinoma, and 2 had mutations in both their AAH and their synchronous adenocarcinoma. In just 1 of the 18 patients was the same K-ras mutation present in the AAHs and adenocarcinoma of the patient. The detection of independent activating point mutations in a cancer-causing gene points to the neoplastic nature of AAH and suggests that glandular neoplasms of the lung arise from a background of field cancerization.

To elucidate the genetic alterations that occur in salivary gland tumors, we screened every autosomal arm (and the X-chromosome) of 29 primary human salivary gland neoplasms (11 pleomorphic adenomas, 10 adenoid cystic carcinomas, 5 mucopidermoid carcinomas, and 3 carcinomas ex-mixed tumors) for allelic loss using 86 microsatellite markers. A minimum of two microsatellite markers were used per chromosomal arm to achieve informativity of at least 60% (excluding X). The pleomorphic adenomas demonstrated few areas of allelic loss; the most prominent chromosomal arm involved was 12q, lost in more than 35% of informative cases. The most significant allelic losses in adenoid cystic carcinoma were 1p, 2p, 6q, 17p, and 20p (>20% of informative cases) and 19q (40% of informative cases). Mucoepidermoid carcinoma showed 50% or greater loss at 2q, 5p, 12p, and 16q. Although losses at 9p, 3p, and 17p are common in squamous cell carcinoma of the head and neck, only the carcinoma ex-mixed tumors demonstrated loss at these loci, consistent with progression to a more aggressive phenotype. Salivary gland tumors display allelic loss patterns different from many other tumor types, suggesting distinct genetic pathways in the progression of these tumors.

Our recent allelic analysis of head and neck squamous cell carcinomas identified a high incidence of chromosomal loss on 13q. To further define an area of minimal loss, we tested 60 primary head and neck squamous cell carcinomas in 59 patients for loss of heterozygosity (LOH) by using 10 polymorphic microsatellite markers spanning the long arm of chromosome 13. We examined the same primary tumors for inactivation of the retinoblastoma (Rb) gene by immunohistochemical analysis of paraffin-embedded specimens. Thirty-one of 60 (52%) tumors demonstrated LOH in at least one 13q marker. Twenty-nine of 31 (94%) lost a portion of 13q that included D13s133, which lies just telomeric to the Rb gene at 13q14.3. However, immunohistochemical staining revealed absence of Rb protein in only 6 of these 31 tumors (19.4%) with LOH. All but one tumor without LOH on 13q displayed normal Rb protein staining. Although Rb may be inactivated by an unusual mechanism in some head and neck squamous cell carcinomas, our data suggest that another tumor suppressor gene locus at 13q14 is likely to be involved in head and neck tumor progression.